Accepted Manuscript

Evaluation of a full-scale wastewater treatment plant upgraded with ozonation and

biological post-treatments: Abatement of micropollutants, formation of transformation
products and oxidation by-products

Marc Bourgin, Birgit Beck, Marc Boehler, Ewa Borowska, Julian Fleiner, Elisabeth
Salhi, Rebekka Teichler, Urs von Gunten, Hansruedi Siegrist, Christa S. McArdell
PII: S0043-1354(17)30873-4
DOI: 10.1016/j.watres.2017.10.036
Reference: WR 13293

To appear in: Water Research

Received Date: 7 August 2017

Revised Date: 8 October 2017
Accepted Date: 17 October 2017

Please cite this article as: Bourgin, M., Beck, B., Boehler, M., Borowska, E., Fleiner, J., Salhi, E.,
Teichler, R., von Gunten, U., Siegrist, H., McArdell, C.S., Evaluation of a full-scale wastewater
treatment plant upgraded with ozonation and biological post-treatments: Abatement of micropollutants,
formation of transformation products and oxidation by-products, Water Research (2017), doi: 10.1016/
j.watres.2017.10.036.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Fate of micropollutants

PT
Biologically
treated water

RI
SC
WWTP
effluent

U
Sand filter (SF)

AN
M
Textile
filter
GACfresh/OZO GACloaded/OZO Fixed bed (FB) Moving bed (MB)

D
TE
GACfresh/BIO EP
Fresh Pre-loaded
C

Fresh
AC

Textile
filter
 ACCEPTED MANUSCRIPT
1 Evaluation of a full-scale wastewater treatment plant upgraded with

2 ozonation and biological post-treatments: Abatement of micropollutants,

3 formation of transformation products and oxidation by-products

4
5 Marc Bourgin1, Birgit Beck1, Marc Boehler1, Ewa Borowska1, Julian Fleiner1, Elisabeth Salhi1, Rebekka

PT
6 Teichler1, Urs von Gunten1,2,3, Hansruedi Siegrist1, Christa S. McArdell1,*

RI
8
1
9 Eawag, Swiss Federal Institute of Aquatic Science and Technology, CH-8600 Dübendorf, Switzerland

SC
2
10 School of Architecture, Civil and Environmental Engineering (ENAC), Ecole Polytechnique Fédérale

11 de Lausanne (EPFL), CH-1015 Lausanne, Switzerland

U
3
12 Institute of Biogeochemistry and Pollutant Dynamics (IBP), ETH Zurich, CH-8092 Zurich, Switzerland
AN
13

14 * Corresponding author: Christa S. McArdell, christa.mcardell@eawag.ch

M

15

16 Key words: wastewater treatment, ozonation, granular activated carbon, transformation products,
D

17 bromate, nitrosamines
TE
C EP
AC
 ACCEPTED MANUSCRIPT
18 ABSTRACT

19 To protect the ecosystem and drinking water resources in Switzerland and in the countries of the

20 downstream catchments, a new Swiss water protection act entered into force in 2016 aiming to reduce

21 the discharge of micropollutants from wastewater treatment plants (WWTPs). As a consequence,

22 selected WWTPs must be upgraded by an advanced treatment for micropollutant abatement with

PT
23 suitable and economic options such as (powdered) activated carbon treatment or ozonation. WWTP

24 Neugut (105’000 people equivalent) was the first WWTP in Switzerland to implement a long-term full-

RI
25 scale ozonation. Differing specific ozone doses in the range of 0.35-0.97 g O3/g DOC were applied to

26 determine the adequate ozone dose to fulfill the requirements of the Swiss water protection act. Based

SC
27 on this assessment, a specific ozone dose of 0.55 g O3/g DOC is recommended at this plant to ensure an

28 average abatement of the twelve selected indicator substances by ≥80% over the whole treatment. A

29
U
monitoring of 550 substances confirmed that this dose was very efficient to abate a broad range of
AN
30 micropollutants by >79% on average. After ozonation, an additional biological post-treatment is

31 required to eliminate possible negative ecotoxicological effects generated during ozonation caused by
M

32 biodegradable ozonation transformation products (OTPs) and oxidation by-products (OBPs). Three
D

33 biological treatments (sand filtration, moving bed, fixed bed) and granular activated carbon (GAC, fresh
TE

34 and pre-loaded) filtration were evaluated as post-treatments after ozonation. In parallel, a fresh GAC

35 filter directly connected to the effluent of the secondary clarifier was assessed. Among the three purely
EP

36 biological post-treatments, the sand filtration performed best in terms of removal of dissolved organic

37 carbon (DOC), assimilable organic carbon (AOC) and total suspended solids (TSS). The fresh activated
C

38 carbon filtration ensured a significant additional micropollutants abatement after ozonation due to
AC

39 sorption. The relative abatement of the indicator substances ranged between 20-89% after 27’000 bed

40 volumes (BV) and was still substantial at 50’000 BV. In an identical GAC filter running in parallel and

41 being fed with the effluent of the secondary clarifier, the elimination was less efficient. Seven primary

42 OTPs (chlorothiazide and six N-oxides) formed during ozonation could be quantified thanks to available

43 reference standards. Their concentration decreased with increasing specific ozone doses with the

44 concomitant formation of other OTPs. The seven OTPs were found to be stable compounds and were

45 not abated in the biological post-treatments. They were sorbed in the fresh GAC filter, but less
 ACCEPTED MANUSCRIPT
46 efficiently than the corresponding parent compounds. Two OBPs, bromate (BrO3-) and N-

47 nitrosodimethylamine (NDMA), were formed during ozonation but did not exceeded 5 µg/L for bromate

48 and 30 ng/L for NDMA at the recommended specific ozone dose of 0.55 g O3/g DOC. NDMA was well

49 abated in all post-treatments (minimum 41% during fixed bed filtration, maximum 83% during fresh

50 GAC filtration), while bromate was very stable as expected.

PT
RI
U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
51 INTRODUCTION

52 More than 100’000 compounds are registered today in Europe (European Chemical Agency 2017),

53 including pharmaceuticals, personal care products, food additives, and industrial chemicals. A large

54 number of them are frequently observed as persistent compounds in surface waters in levels of ng/L-

55 µg/L (Petrie et al. 2014, Richardson and Kimura 2016). Therefore, sensitive species in the aquatic

PT
56 ecosystem may be adversely affected by, e.g., a damage of the nervous systems, feminization and

57 disruption of reproduction, or inhibition of photosynthesis (Stalter et al. 2013). Furthermore,

RI
58 micropollutants in surface waters and in groundwaters can contaminate the drinking water resources

59 (Lopez et al. 2015).

SC
60 Many micropollutants observed in water bodies originate from WWTP discharges, where polar

61 micropollutants are generally poorly abated by conventional treatments (Loos et al. 2013), as WWTPs

62
U
are not designed to eliminate micropollutants. The task of conventional WWTPs is to remove nutrients
AN
63 (inorganic nitrogen and total phosphorus), suspended solids and to reduce the biological and chemical

64 oxygen demand (BOD5 and COD, respectively) (Ternes et al. 2017).

M

65 In Switzerland, a new water protection act entered into force as of January 2016 with the objective to
D

66 improve surface water quality by reducing the load of micropollutants from WWTPs. Within the next
TE

67 twenty years, about 100 of the 700 existing Swiss WWTPs will have to implement an advanced

68 treatment step for micropollutant abatement, following four criteria: (i) large WWTPs to significantly
EP

69 reduce the load; (ii) WWTPs in the catchment of lakes; (iii) WWTPs at rivers with a high fraction of

70 wastewater to protect ecosystems; and (iv) WWTPs at surface waters which serve as drinking water
C

71 resources.
AC

72 The aim is to remove organic micropollutants by 80% on average over the whole WWTP from influent

73 to effluent (Eggen et al. 2014). To ensure the efficiency of the upgraded wastewater treatment, the Swiss

74 Federal Office of Environment (FOEN) has established a shortlist of 12 indicator substances (SI, Table

75 S1), which must be abated on average by 80% (FOEN 2017). The 12 compounds were specifically

76 chosen as non-easily biodegradable substances and are therefore normally not well removed during

77 conventional wastewater treatments.

 ACCEPTED MANUSCRIPT
78 Adsorption by powdered activated carbon (PAC) and ozonation are currently the recommended

79 technologies that lead cost-efficiently to a significant abatement of the micropollutant load and toxicity

80 (Hollender et al. 2009, Logar et al. 2014, Margot et al. 2013, Schindler Wildhaber et al. 2015). Besides,

81 a treatment with granular activated carbon filtration has been recently considered as a potential

82 alternative and is already applied on a full-scale WWTP in Germany (Benstoem et al. 2017).

PT
83 Ozone reacts quickly with electron-rich moieties, e.g., olefins, tertiary amines, thioethers and activated

84 aromatics (von Sonntag and von Gunten 2012), while hydroxyl radicals (•OH), which are formed from

RI
85 self-decomposition of ozone (Hoigné and Bader 1975, Staehelin et al. 1984, Staehelin and Hoigné

86 1982), react additionally with alkanes, amides and non-activated aromatic compounds (Buxton et al.

SC
87 1988). If the mechanism and the kinetics of the reactions of a substance or moiety with ozone are

88 known, the possible OTPs and the rate of abatement of a substance can in principle be predicted (Hübner

89
U
et al. 2015, Lee et al. 2017, Lee and von Gunten 2016, von Sonntag and von Gunten 2012). Even though
AN
90 there is a lot of mechanistic information available, there are still significant knowledge gaps on certain

91 functional groups.
M

92 Furthermore, ozonation may yield oxidation/disinfection by-products (OBPs) by reaction of ozone and
D

93 hydroxyl radicals with the different components of the water matrix. Some OBPs such as BrO3- and
TE

94 nitrosamines, e.g., N-nitrosodimethylamine (NDMA), are known to be (potential) human carcinogens

95 (Kurokawa et al. 1990, Tricker and Preussmann 1991). Therefore, the formation of these OBPs needs to
EP

96 be minimized during ozonation. In Switzerland, the formation of BrO3- and NDMA during ozonation

97 needs to be assessed according to a recently introduced test procedure to evaluate the treatability of
C

98 wastewater with ozone (Schindler Wildhaber et al. 2015). Bromate is formed from the successive
AC

99 oxidative reactions of bromide (Haag and Hoigné 1983, von Gunten and Hoigné 1994) present in water

100 bodies and wastewaters in µg/L-mg/L levels (Soltermann et al. 2016). The main bromide sources in

101 Switzerland are from wastewaters discharged from municipal waste incinerators, industrial landfills and

102 chemical industries (Soltermann et al. 2016). A bromate drinking water standard/guideline value for

103 BrO3- was set to 10 µg/L (Commission Directive 2003, U.S. EPA 2006, WHO 2011) and acute and

104 chronic environmental quality standards (EQS) of 50 µg/L were proposed by the Ecotox Centre Eawag-

105 EPFL (Ecotox Center 2017). NDMA is formed during ozonation of precursors such as hydrazines and
 ACCEPTED MANUSCRIPT
106 sulfamides (yield >50%) (Kosaka et al. 2009, Krasner et al. 2013, Lim et al. 2016, Schmidt and Brauch

107 2008, von Gunten et al. 2010). The NDMA guideline value in drinking water was set to 100 ng/L by

108 WHO (2008) and even lower values of 10 ng/L are proposed in countries such as the USA and Germany

109 (California Department of Public Health 2010, Massachusetts Department of Environmental Protection

110 2004, Planas et al. 2008).

PT
111 Some toxicity endpoints may increase after ozonation due to the formation of aldehydes and organic

112 acids. It has been recently concluded that OBPs are toxicologically probably more relevant than OTPs

RI
113 (Lee and von Gunten 2016). However, toxicity was found to decrease again after biological post-

114 treatments (Stalter et al. 2010a, Stalter et al. 2010b), because of the biodegradability of toxic OBPs and

SC
115 OTPs. Therefore, a biological post-treatment is required with the goal to eliminate biodegradable, toxic

116 OTPs and OBPs leading to negative ecotoxicological effects.

117
U
The present study aims at evaluating a full-scale wastewater treatment with installed ozonation and post-
AN
118 treatment. To this end, three purely biological post-treatments (a sand filter, a moving bed reactor, and a

119 fixed bed reactor) were investigated as well as GAC filtration with additional sorption capacity (a fresh
M

120 and a pre-loaded GAC filter). Additionally, a fresh GAC filter was run in parallel directly with the
D

121 secondary effluent for comparison. The effects of the differing treatment steps were assessed on the
TE

122 chemical (current paper) and biological (Kienle et al. in preparation) wastewater quality. Additionally,

123 non-target peaks across the wastewater treatment train were characterized (Schollee et al. 2017). Here,
EP

124 the abatement of 43 micropollutants frequently detected in wastewater, including the indicator

125 substances, was first investigated at various ozone doses to determine the required specific ozone dose to
C

126 abate the indicator substances by 80% on average, as requested by the new Swiss water protection act.
AC

127 Thereafter, the abatement of micropollutants from an exhaustive list of 550 substances was studied at the

128 optimum specific ozone dose for an overall evaluation of the advanced treatment. With a target and

129 suspect screening approach, the formation of known OTPs and the OBPs BrO3- and NDMA during

130 ozonation was investigated. The fate of micropollutants, OTPs and OBPs over the differing biological

131 post-treatments and in the GAC filters was studied and compared.
 ACCEPTED MANUSCRIPT
132 MATERIAL AND METHODS

133 Wastewater treatment plant

134 Neugut WWTP (Dübendorf, Switzerland) (Fig. S1, Text S1) treats wastewater of about 50’000 persons

135 and 105’000 people equivalent (maximum capacity of 150’000 people equivalent) with an approximate

136 industrial contribution of 50% in COD load (mainly from food industries). The flow in the treatment

PT
137 plant ranges between 13’000 and 57’000 m3/day with an average of 19’000 m3/day at dry weather

138 conditions. The wastewater treatment originally consisted of a primary clarifier, a conventional activated

RI
139 sludge treatment with a secondary clarifier and a sand filtration. The average sludge retention time in the

140 conventional biological treatment is about 13 days and the hydraulic retention time (HRT) is 18 h

SC
141 (without settling tank). Dissolved organic carbon (DOC) after the conventional biological treatment was

142 3.5 – 6 mg/L and nitrite below 0.04 mg N/L; other quality parameters are given in Table S2). Although

143
U
the industrial contribution is relatively high in this WWTP, the quality of the secondary effluent is
AN
144 similar to typical Swiss domestic wastewater (Schindler Wildhaber et al. 2015, Soltermann et al. 2016).

145 In 2014, an ozonation reactor (V = 530 m3, divided in 6 reactors in series) was installed between the
M

146 secondary clarifiers and the four sand filters (4x31.5 m3 operated in parallel, quartz sand, 0.7-2.0 mm
D

147 granular size with an average empty bed contact time (EBCT) of 10 min). Ozone is produced with two
TE

148 ozone generators (OZAT-CFV5, Ozonia, Dübendorf, Switzerland) from oxygen fed by a liquid oxygen

149 tank. The ozone reactor has an average HRT of 43 minutes at dry weather conditions and of 13 minutes
EP

150 at maximum flow. Furthermore, 4 post-treatments, i.e., a moving bed (MB, 300 L, average EBCT 21

151 min), a fixed bed (FB, 220 L, average EBCT 25 min), a fresh GAC filter (GACfresh/OZO, 77 L, i.e. 34.65
C

152 kg Cyclecarb 401, density 450 g/L, 0.425-2.36 mm granular size, average EBCT 14.5 min) and a pre-
AC

153 loaded GAC filter (GACloaded/OZO, 85 L, i.e. 40.8 kg Norit GCN 830, density 480 g/L, 1.4-2.36 mm

154 granular size, average EBCT 18 min) were installed in pilot-scale for a comparison with sand filter, run

155 flow proportionally at dry weather water flow. An identical fresh GAC filter (GACfresh/bio, 77 L,

156 Cyclecarb 401, 0.425-2.36 mm granular size, average EBCT 14 min) was installed directly after the

157 secondary clarifier. More details about the differing reactors and the operational parameters are given in

158 Text S1.

159 Wastewater sampling and chemical analyses

 ACCEPTED MANUSCRIPT
160 Wastewater sampling

161 Samples were taken at the effluent of the primary clarifier (WWTP-INF), of the secondary clarifier

162 (BIO-EFF), of the ozonation reactor (OZO-EFF) and of the post-treatments mentioned above (SF-EFF,

163 MB-EFF, FB-EFF, GACfresh/bio-EFF, GACfresh/OZO-EFF and GACloaded/OZO-EFF). On 37 sampling days

164 over a 10-month period (September 2014-June 2015), samples (1 L, 24-h flow proportional) were

PT
165 collected from all sampling points in intervals from 1 day at the beginning to 5 weeks at the end. Note

166 that WWTP-INF was always sampled 1 day before the others to account for the HRT of the plant.

RI
167 Four sampling campaigns (September/October 2014) were conducted collecting 24-h flow proportional

SC
168 samples (1 L) at WWTP-INF, BIO-EFF, OZO-EFF and SF-EFF on 3 consecutive dry days at differing

169 ozone doses of 2.0 (October 3-5), 3.0 (September 24-26), 4.0 (September 11-13) and 4.9 mg O3/L

U
170 (September 15-17), corresponding to specific ozone doses of 0.35±0.02, 0.54±0.04, 0.67±0.02 and
AN
171 0.97±0.07 g O3/g DOC, respectively. Thereafter, ozonation was run at 2.7 mg/L ozone (about 0.54 g

172 O3/g DOC), the ozone dose recommended for continuous full-scale application at the WWTP Neugut. In
M

173 February and March 2015, two sampling campaigns (48-h flow proportional samples for all sampling

174 points) were carried out to screen for 550 environmentally relevant micropollutants.
D

175 The samples were collected at 4°C and were maintained at this temperature until analyses. If chemical
TE

176 analyses could not be carried out within a week, the samples were frozen at -20°C and thawed just

177 before sample preparation.

EP

178 Quantification of 51 micropollutants and OTPs using an automatized extraction (online SPE-LC-HRMS,
C

179 method I)
AC

180 The concentration of 43 micropollutants (acesulfame, aliskiren ,amisulpride, atenolol, atenolol acid,

181 azithromycin, benzotriazole, bezafibrate, candesartan, carbamazepine, carbendazim, cetirizine,

182 citalopram, clarithromycin, DEET, diclofenac, diuron, eprosartan, fexofenadine, gabapentin,

183 hydrochlorothiazide, iopromide, irbesartan, lamotrigine, levetiracetam, losartan, mecoprop, metoprolol,

184 methylbenzotriazole, N4-acetyl-sulfamethoxazole, oxazepam, phenazone, primidone, ranitidine,

185 sucralose, sulfamethoxazole, telmisartan, tramadol, triclosan, trimethoprim, valsartan, valsartan acid,

186 venlafaxine, including parent compounds and metabolites, Table S5) and eight OTPs with reference
 ACCEPTED MANUSCRIPT
187 standards (4’-/5-hydroxydiclofenac, amisulpride N-oxide, cetirizine N-oxide, clarithromycin N-oxide,

188 fexofenadine N-oxide, chlorothiazide, tramadol N-oxide and venlafaxine N-oxide) was determined

189 according to Jeon et al. (2013) (details of method I are presented in Text S2, SI). Shortly, samples were

190 filtered through 2 glass microfiber filters, i.e., GF/D (top, 2.7 µm pore size, Whatman) and GF/F

191 (bottom, 0.7 µm pore size, Whatman). After dilution and spiking with internal standards (IS), the

PT
192 samples were extracted by online solid-phase extraction with a combination of four adsorbents (Oasis

193 HLB and a Strata-X-AW/Strata-X-CW/ENV+ mixture). The extracts were directly analyzed by liquid

RI
194 chromatography (Atlantis T3 column, Waters), coupled to high-resolution mass spectrometry (Q-

195 Exactive, ThermoScientific). Details on quality control measurements (limit of quantification and

SC
196 recoveries) are given in Table S5.

197 Additionally to the eight OTPs mentioned above, other OTPs, presented in the literature (Table S6),

198
U
without reference standards were screened with their exact mass of the protonated or deprotonated
AN
199 molecular ion. MS/MS spectra from fragmentation did not show high enough intensities for a

200 confirmation.
M

201 Quantification of 550 micropollutants using a manual extraction (offline SPE-LC-HRMS, method II)
D
TE

202 The fate of 550 environmentally-relevant chemicals was investigated with an offline SPE-LC-HRMS

203 method, adapted from Kern et al. (2009) and described in Text S3, SI. This list included 240
EP

204 pharmaceuticals, 234 pesticides, 38 industrial chemicals and 60 biological TPs. Their physical-chemical

205 properties cover a wide range of molecular weight (120-915 g/mol), log Kow (-2.2 – 5.7), and charge at
C

206 circumneutral pH (50% neutral, 50% ionic). In brief, 1 L of sample was filtered (as described above) and
AC

207 spiked with IS. The samples were then extracted through a multilayered solid-phase cartridge, consisting

208 of ENVI-Carb activated carbon (bottom), Strata-X-AW/Strata-X-CW/ENV+ mixture (middle layer), and

209 Oasis HLB (top layer). Chemical analyses were carried out with the same LC-HRMS system as

210 described above.

211 Determination of bromide and bromate

 ACCEPTED MANUSCRIPT
212 For the quantification of bromide and bromate, 1 mL samples (24- and 48-h flow proportional,

213 September 2014-June 2015, n=30) were filtered with a 13-mm syringe filter (regenerated cellulose

214 membrane, 0.45µm porosity, Infochroma AG), diluted 100 times with ultrapure water and stored at

215 +4°C until analyses. Bromide and bromate were separated by capillary ion chromatography (Thermo

216 Dionex ICS-4000) using an anion-exchange IC column (IonPac AS19-4µm, Dionex) and detected by a

PT
217 modified tandem mass spectrometry (Thermo TSQ-Vantage) to reach lower detection limit (Shah et al.

218 2015). The quantification limits of bromide and bromate in wastewater were 25 µg/L and 2 µg/L,

RI
219 respectively, for the 100-fold dilution (0.25 and 0.02 µg/L, respectively, in ultrapure water without

220 dilution).

SC
221 Determination of N-nitrosodimethylamine (NDMA) by LC-HRMS/MS

U
AN
222 The concentrations of NDMA were determined in several samples collected flow proportionally in

223 September-October 2014 (n=2, 24-h composite; n=3 72-h composite) and in February-April 2015 (n=3,
M

224 48-h composite) with LC-HRMS/MS according to a method described previously (Krauss and Hollender

225 2008), obtaining a limit of quantification at 5 ng/L (Text S5).

D
TE

226 Determination of second-order rate constants

227 For some relevant compounds without available kinetic data, apparent second-order rate constants at pH
EP

228 7 (kO3,pH7) were determined: 1.5
105 M-1 s-1 for amisulpride, 4.9
105 M-1 s-1 for eprosartan, 2.1
105
229 M-1 s-1 for losartan, 1.2
105 M-1 s-1 for telmisartan, 560 M-1 s-1 for candesartan, and 24 M-1 s-1 for
230 irbesartan (See Text S4 for details). These and other rate constants from literature are compiled in Table
C

231 S4.
AC
 ACCEPTED MANUSCRIPT
232 RESULTS AND DISCUSSION

233 Micropollutant abatement.

234 Fate of micropollutants during ozonation and over the whole wastewater treatment

235 The abatement of 43 wastewater relevant micropollutants (Table 1, parent compounds and metabolites)

PT
236 in biological treatment with conventional activated sludge (BIO) and during ozonation (OZO) with 4

237 specific ozone doses is presented in Table 1 (concentrations can be found in Table S5).

RI
238 During the biological treatment, the abatements were relatively constant for single compounds over the

239 sampling period (n=12), but varied considerably between -131% and >99% among them (SI, Table S8).

SC
240 A limited number of compounds (acesulfame, bezafibrate, DEET, eprosartan, levetiracetam, N4-acetyl-

241 sulfamethoxazole, triclosan, trimethoprim and valsartan) were almost completely eliminated in

242
U
biological treatment (relative abatement >80%). Half of the compounds (n=22) were not significantly
AN
243 eliminated (relative abatement <20%). Some investigated compounds, e.g., lamotrigine, phenazone and

244 the TP valsartan acid, were even significantly formed during this treatment step (negative elimination or
M

245 detection of the compound only after BIO). For lamotrigine, this is in agreement with recent studies
D

246 (Bollmann et al. 2016, Zonja et al. 2016) which showed that non-conjugated and conjugated forms
TE

247 (mainly as glucuronide) enter the WWTPs. In conventional biological treatment, conjugated forms were

248 cleaved to lamotrigine, which is relatively stable. Therefore, its concentration significantly increases,
EP

249 typically from about 300 to 800 ng/L (Table S5). Phenazone was also found to be formed from other

250 pharmaceuticals (from metamizole or human metabolites) during biological treatment (Gros et al. 2010,
C

251 Kovalova et al. 2012, Prasse et al. 2010), and valsartan acid was identified as a biodegradation product
AC

252 of valsartan (Kern et al. 2010).

253 During ozonation, the extent of abatement increased with increasing ozone doses (Table 1 and Fig. S4).

254 At the lowest specific ozone dose of 0.35 g O3/g DOC, 20 out of the 43 substances were abated by less

255 than 80%. However, this number decreased significantly to 13, 11 and 5 with higher specific ozone

256 doses of 0.54 g O3/g DOC, 0.67 g O3/g DOC and 0.97 g O3/g DOC, respectively. Substances with high

257 reactivity with ozone at circumneutral pH (Table S4), e.g., diclofenac, carbamazepine,

258 hydrochlorothiazide, sulfamethoxazole and tertiary amines such as cetirizine, fexofenadine and
 ACCEPTED MANUSCRIPT
259 tramadol, were very well abated (>80%) even at 0.35 g O3/g DOC. For higher specific ozone doses,

260 these compounds were abated by >93% and not detected anymore (<LOQ). Compounds with low

261 apparent second-order rate constant for the reaction with ozone at circumneutral pH (i.e., kO3,pH7 <10 M-1

262 s-1), for which the reaction with •OH is the predominant abatement pathway, e.g., iopromide, lamotrigine

263 and sucralose, were only slightly abated (<30%) at a specific ozone dose of 0.35 g O3/g DOC. The

PT
264 concentration of these compounds was reduced by less than 71% at the highest specific ozone dose (0.97

265 g O3/g DOC).

RI
266 Other studies, presented in Table S9, showed similar abatements for comparable specific ozone doses

267 (Kovalova et al. 2013, Lee et al. 2013, Reungoat et al. 2010, Ternes et al. 2017) or slightly lower

SC
268 (Hollender et al. 2009, Margot et al. 2013). In the latter two studies, the abatement of compounds

269 predominantly reacting with •OH (methylbenzotriazole, benzotriazole, gabapentine, iopromide,

270
U
irbesartan, mecoprop, oxazepam, primidone and valsartan) was particularly low. These differences may
AN
271 be attributed to varying wastewater qualities, e.g., a differing composition of the dissolved organic

272 matter leading to a lower •OH formation (Schindler Wildhaber et al. 2015) or a higher alkalinity (HCO3-
M

273 being a •OH scavenger (Buxton and Elliot 1986) in these wastewaters. In the study by Margot et al.
D

274 (2013), the presence of nitrite in high concentrations (average value of 0.4 mg NO2-N/L, compared to
TE

275 0.01-0.04 mg NO2-N/L level in other studies) explains at least partially the lower abatements. Nitrite

276 scavenges ozone rapidly in a ratio of about 1 g O3 per g NO2- (1:1 molar ratio), therefore, a specific
EP

277 ozone dose of 0.83 g O3/g DOC corresponds to an effective specific dose of 0.63 g O3/g DOC (DOC 7.3

278 mg /L) at 1.4 mg NO2-/L (i.e., 0.4 mg NO2-N/L).

C

279 Over the entire treatment, including biological treatment and ozonation (BIO+OZO), most of the
AC

280 compounds were abated by >80% at a specific ozone dose of 0.35 g O3/g DOC, except 10 compounds

281 (Table 1). At specific ozone doses of 0.54 g O3/g DOC, 0.67 g O3/g DOC or 0.97 g O3/g DOC, 6, 7 or 2

282 compounds, respectively, were not abated >80%.

283 To comply with the guidelines established by the Swiss authorities for the evaluation of advanced

284 wastewater treatment, 12 indicator substances need to be eliminated by 80% on average over the whole

285 treatment chain (marked in Table 1, see Table S1 and Table S8 for more details). These 12 indicator

286 substances were abated by 2-44% during biological treatment, except benzotriazole (64±4%).
 ACCEPTED MANUSCRIPT
287 Benzotriazole is often poorly abated in biological treatment (Herzog et al. 2014, Liu et al. 2011), except

288 in some specific WWTPs (Mazioti et al. 2015), for unknown reasons. The average abatement increased

289 to 85 ± 3% when the biological treatment was followed by a low specific ozone dose (0.35 g O3/g DOC)

290 and even up to >94% when the specific ozone dose was 0.54 g O3/g DOC. Note that these 12 substances

291 are not abated in sand filtration, as will be discussed in more detail below.

PT
292 Considering these results, a combination of biological treatment and a specific ozone dose of 0.40 g O3/g

293 DOC (1.5-2.5 mg O3/L) is sufficient to reach an average abatement of >80% of the 12 indicator

RI
294 substances as requested by the Swiss water protection act (FOEN 2017). However, to compensate a

295 potential nitrite peak concentration of up to 0.7 mg NO2-/L (i.e., 0.2 mg NO2-N/L, while the usual nitrite

SC
296 concentration at WWTP Neugut is 0.02-0.04 mg NO2-N/L), a specific ozone dose of 0.55 g O3/g DOC

297 (2.0-3.3 mg O3/L) is recommended for routine operation at this site.

298
U
Two sampling campaigns (48-h flow proportional samples) were carried out to investigate the abatement
AN
299 of an exhaustive list of 550 micropollutants over the whole treatment chain at a specific ozone dose of

300 0.55 g O3/g DOC (Fig. 1, Fig. S5, Table S7). In WWTP-INF, 175 compounds were found (>LOQ) with
M

301 a total concentration of 371 µg/L. Most abundant were pharmaceuticals (133 µg/L), caffeine (130 µg/L)
D

302 and food additives (85 µg/L) in the influent, while pesticides and biocides were detected only in very
TE

303 low concentrations (ca. 1 µg/L). Being aware that the sampling was done in winter, it may not be

304 representative for the whole year. In biological treatment, the total concentration of target compounds
EP

305 decreased by 85% to 54 µg/L, although the number of quantified compounds increased to 204

306 substances. This increase can be explained by the formation of biological transformation products that
C

307 were included in the target list, and the lower LOQ in BIO-EFF compared to WWTP-INF due to
AC

308 differing dilutions used for the analyses and due to higher matrix effects in WWTP-INF (see Text S3).

309 The 4 most abundant compounds in WWTP-INF, i.e., caffeine, cyclamate, metformin and paracetamol

310 (total concentration 260 µg/L), were eliminated by 99%. Without the contribution of these 4 compounds,

311 the abatement of the remaining micropollutant load drops from 85% to 54% during the biological

312 treatment. After ozonation, the total concentration was still 23 µg/L, corresponding to an abatement of

313 the micropollutants load by 58%. In sand filtration, no significant elimination was observed, and over

314 the whole plant, the total load decreased by 94%. Five compounds (acesulfame, sucralose, metformin,
 ACCEPTED MANUSCRIPT
315 iopromide and tris(2-chloropropyl)phosphate) were found in concentrations >0.9 µg/L in OZO-EFF. All

316 were poorly abated during ozonation (<50%). Alternatively, the abatement can be considered for each

317 single substance and an average can be calculated. As a result, the average abatement of all substances

318 (taking a zero abatement for all substances with negative abatement into account) over the biological

319 treatment is 34% (median 22%), during ozonation >67% (median >74%), and over the whole WWTP,

PT
320 with and without sand filter, >79% (median >89%) (Fig. 1). Due to the fact that 64 substances (out of

321 147) were not detected anymore after ozonation, only a minimal abatement relative to the LOQ could be

RI
322 calculated in these cases.

U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

323 Table 1. Relative abatements (%) of 43 micropollutants during the differing steps of wastewater treatment. BIO: conventional biological treatment;
324 OZO: ozonation at 4 differing specific ozone doses (0.35-0.97 g O3/g DOC); BIO+OZO: overall from the inlet of conventional biological treatment to the
325 outlet of the ozonation reactor. Average values and standard deviations were calculated for 3 consecutive dry weather days (n=3). Abatements with a
326 “>” sign indicate that the concentration at the outlet of the reactor was below LOQ and the given % abatement corresponds to the minimum abatement

PT
327 which was calculated using the compound’s LOQ.
0.35±0.02 g O3/g DOC 0.54±0.05 g O3/g DOC 0.67±0.03 g O3/g DOC 0.97±0.07 g O3/g DOC

RI
BIO OZO BIO+OZO BIO OZO BIO+OZO BIO OZO BIO+OZO BIO OZO
Acesulfame 95 ± 1 39 ± 5 97 ± 1 90 ± 8 59 ± 9 96 ± 3 95 ± 1 70 ± 8 98 ± 1 88 ± 5 >90 ± 4

SC
Aliskiren 25 ± 4 >84 ± 1 >88 ± 1 23 ± 22 >93 ± 1 >95 ± 1 7 ± 12 >81 ± 1 >82 ± 3 8 ± 11 >92 ± 3
Amisulpride† 2 ± 14 91 ± 1 91 ± 2 3 ± 16 >98 ± 1 >98 ± 1 -1 ± 13 >95 ± 1 >95 ± 1 3±8 >98 ± 1
Atenolol 75 ± 3 70 ± 2 92 ± 1 70 ± 5 92 ± 3 98 ± 1 72 ± 5 93 ± 5 98 ± 1 71 ± 4 >93 ± 1

U
Atenolol acid 76 ± 2 70 ± 3 93 ± 1 72 ± 3 91 ± 3 98 ± 1 67 ± 4 >92 ± 4 >97 ± 1 69 ± 3 >98 ± 1
Azithromycin 8 ± 28 >90 ± 1 >90 ± 3 n.m.2 >54 ± 3 n.m. 25 ± 14 >88 ± 2 >91 ± 3 23 ± 15 >95 ± 1

AN
Benzotriazole† 66 ± 3 52 ± 1 83 ± 1 62 ± 7 74 ± 3 90 ± 3 63 ± 6 80 ± 7 93 ± 2 64 ± 3 91 ± 4
Bezafibrate 96 ± 1 >62 ± 5 >98 ± 1 >95 ± 2 >75 ± 1 >99 ± 1 >93 ± 2 n.m.1 >97 ± 1 >94 ± 1 >24∆
Candesartan† 0 ± 16 63 ± 1 63 ± 6 1 ± 12 82 ± 3 82 ± 5 -24 ± 20 85 ± 6 81 ± 6 -17 ± 19 94 ± 3

M
Carbamazepine† -16 ± 12 95 ± 1 94 ± 1 -14 ± 19 >98 ± 1 >98 ± 1 -33 ± 23 >98 ± 1 >98 ± 1 -24 ± 8 >98 ± 1
# #
Carbendazim 8 ± 18 82 ± 2 83 ± 5 -1 ± 23 >94 ± 3 >94 ± 4 7 ± 16 >93 ± 5 >89 ± 8 -28 ± 50 >89 ± 4

D
Cetirizine 6 ± 12 92 ± 1 92 ± 2 -9 ± 12 >93 ± 1 >92 ± 1 -25 ± 25 >95 ± 1 >94 ± 1 -19 ± 10 >95 ± 1
# #

TE
Citalopram† 5±4 91 ± 3 89 ± 4 4±4 >96 ± 3 >96 ± 3 -8 ± 2 >94 ± 1 >93 ± 1 4±8 >97 ± 1
Clarithromycin† 51 ± 2 93 ± 2 97 ± 1 28 ± 19 >95 ± 1 >96 ± 1 45 ± 6 >94 ± 1 >97 ± 1 52 ± 8 >97 ± 1
DEET 96 ± 1 49 ± 9 98 ± 1 >90 ± 3 65 ± 1 >95 ± 2 97 ± 1 52 ± 18 99 ± 1 97 ± 1 70 ± 15
EP
Diclofenac† 22 ± 8 96 ± 1 97 ± 1 23 ± 8 100 ± 1 100 ± 1 11 ± 10 >99 ± 1 >99 ± 1 13 ± 18 >99 ± 1
#
Diuron 0 ± 30 64 ± 2 64 ± 10 -7 ± 7 >84 ± 8 >83 ± 10 -5 ± 14 >84 ± 6 >84 ± 4 -40 ± 72 >90 ± 6
Eprosartan 98 ± 1 >66 ± 3 >99 ± 1 >93 ± 2 n.m.1 >97 ± 1 >97 ± 1 >67 ± 7 >99 ± 1 >88 ± 5 n.m.1
C

Fexofenadine 13 ± 16 83 ± 3 85 ± 6 9±4 >94 ± 1 >95 ± 1 12 ± 12 >96 ± 3 >97 ± 2 -1 ± 12 >91 ± 1

AC

Gabapentin 44 ± 10 44 ± 4 69 ± 6 43 ± 6 55 ± 4 75 ± 1 37 ± 5 63 ± 10 77 ± 6 44 ± 5 75 ± 6
Hydrochlorothiazide† 9 ± 13 86 ± 2 87 ± 4 13 ± 9 >98 ± 2 >98 ± 2 -2 ± 17 >97 ± 3 >97 ± 2 1 ± 10 >99 ± 1
# #
Iopromide 72 ± 5 28 ± 1 80 ± 4 70 ± 8 43 ± 3 83 ± 5 53 ± 20 53 ± 6 78 ± 8 64 ± 8 65 ± 1
Irbesartan† 17 ± 18 57 ± 1 64 ± 8 15 ± 15 75 ± 3 78 ± 5 2 ± 20 79 ± 4 79 ± 6 0 ± 25 89 ± 4
Lamotrigine -125 ± 31 37 ± 1 -42 ± 18 -128 ± 10 50 ± 2 -15 ± 4 -140 ± 30 57 ± 7 -3 ± 21 -132 ± 5 71 ± 6
Levetiracetam 99 ± 1 >66 ± 34 100 ± 1 98 ± 1 >43 ± 36 >99 ± 1 >99 ± 1 n.m.1 100 ± 1 >99 ± 1 n.m.1
Losartan 71 ± 2 93 ± 1 98 ± 1 68 ± 5 >98 ± 1 >99 ± 1 66 ± 5 >98 ± 1 >99 ± 1 67 ± 4 >98 ± 1
 ACCEPTED MANUSCRIPT

# # # # #
Mecoprop -5 ± 2 59 ± 2 56 ± 1 14 ± 7 77 ± 2 81 ± 4 5 ± 30 82 ± 7 84 ± 5 9 ± 14 >91 ± 3
Metoprolol† 41 ± 7 75 ± 2 85 ± 3 38 ± 5 94 ± 3 96 ± 2 37 ± 8 94 ± 5 96 ± 3 38 ± 4 >99 ± 1
Methylbenzotriazole†,* 45 ± 22 66 ± 3 81 ± 9 61 ± 4 89 ± 4 96 ± 2 2 ± 22 90 ± 1 90 ± 7 28 ± 12 98 ± 1
N4-Acetyl-Sulfamethoxazole >98 ± 1 n.m.1 >99 ± 1 >98 ± 1 n.m.1 >99 ± 1 >98 ± 1 n.m. >99 ± 1 >95 ± 2 n.m.1

PT
Oxazepam 5 ± 17 55 ± 1 57 ± 7 10 ± 10 73 ± 3 75 ± 5 -13 ± 22 77 ± 5 74 ± 6 -10 ± 9 88 ± 4
Phenazone n.m.2 >91 ± 1 n.m.1 n.m.2 >92 ± 1 n.m.1 n.m.2 >92 ± 1 n.m.1 n.m.2 >81 ± 4

RI
Primidone 19 ± 12 48 ± 3 58 ± 5 15 ± 17 66 ± 1 72 ± 6 14 ± 14 70 ± 10 75 ± 7 -4 ± 8 86 ± 5
Ranitidine 82 ± 1 >74 ± 2 >95 ± 1 79 ± 4 >80 ± 4 >96 ± 1 66 ± 16 >71 ± 2 >90 ± 5 77 ± 2 >53 ± 19
Sucralose 9 ± 20 21 ± 2 28 ± 14 13 ± 13 27 ± 1 36 ± 10 -9 ± 17 38 ± 4 32 ± 12 -4 ± 13 46 ± 4

SC
Sulfamethoxazole 55 ± 3 85 ± 2 93 ± 1 46 ± 6 >97 ± 1 >98 ± 1 31 ± 6 >97 ± 1 >98 ± 1 34 ± 2 >95 ± 3
Telmisartan 15 ± 3 66 ± 1 71 ± 1 3 ± 16 86 ± 3 86 ± 4 -24 ± 27 88 ± 5 84 ± 8 -7 ± 18 >94 ± 1
Tramadol** 0 ± 11 91 ± 1 90 ± 3 2±7 98 ± 1 98 ± 1 -5 ± 11 >98 ± 1 >98 ± 1 -6 ± 10 >98 ± 1

U
Triclosan n.m.1 n.m.1 n.m.1 n.m.1 n.m.1 n.m.1 n.m.1 >56 ± 1 n.m.1 n.m.1 n.m.1

AN
Trimethoprim >89 ± 4 >59 ± 6 >95 ± 1 85 ± 6 >61 ± 15 >95 ± 1 75 ± 6 >58 ± 4 >90 ± 1 83 ± 3 >79 ± 3
Valsartan 94 ± 1 52 ± 3 97 ± 1 93 ± 2 69 ± 3 98 ± 1 96 ± 1 68 ± 8 99 ± 1 93 ± 1 >86 ± 3
Valsartan acid n.m.2 49 ± 1 n.m.1 n.m.2 67 ± 1 n.m.1 n.m.2 71 ± 5 n.m.1 n.m.2 76 ± 8

M
Venlafaxine† 4 ± 15 84 ± 2 84 ± 5 9 ± 10 >97 ± 2 >97 ± 2 -12 ± 17 >96 ± 3 >96 ± 2 1±5 >96 ± 1
Average of the 12 indicator 17 ± 8 79 ± 1 85 ± 3 20 ± 7 >91 ± 2 >94 ± 2 7±8 >92 ± 3 >93 ± 2 14 ± 9 >96 ± 1
substances†

D
328 n.m.1: abatement not measurable because the concentrations at the inlet and at the outlet of the reactor were both <LOQ.
329 n.m.2: abatement not measurable because the concentration at the inlet is <LOQ; abatement has a negative value.

TE
330 *: methylbenzotriazole represents the sum of 4- and 5-methylbenzotriazole, which co-elute under the analytical conditions.
331 **: tramadol represents the sum of tramadol and o-desmethyl-venlafaxine, which are isobaric compounds that co-eluted under the analytical conditions.
332 #: abatement was determined for only 2 samples due to high fluctuations of concentrations in one sample.
∆
EP
333 : abatement was determined for only 1 sample due to high fluctuations of concentrations in two samples.
334 †: the 12 indicator substances (bold in the table above) are amisulpride, benzotriazole, candesartan, carbamazepine, citalopram, clarithromycin, diclofenac, hydrochlorothiazide, irbesartan,
335 methylbenzotriazole, metoprolol and venlafaxine.
C
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

336

337 Fig. 1. Sum of concentrations of micropollutants at various WWTP sampling points (n=2, 48-h
338 flow proportional samples, specific ozone dose 0.55 g O3/g DOC, analyzed with method II; for
339 sampling points see Fig. S1). *: abatement of total concentration relative to WWTP-INF (effluent
340 of primary clarifier); **: abatement relative to BIO-EFF (effluent of conventional biological
341 treatment); ***: abatement relative to OZO-EFF (effluent of ozonation reactor). Inset: Box-
342 whisker plot depicting the distribution of calculated abatements for all substances (minimum, 25th
343 percentile, median, 75th percentile, maximum and ( ) average) of micropollutants (analyzed with
344 method II) during all treatment steps. Because nearly half of the compounds were not detected
345 anymore after ozonation, the values given for OZO, BIO+OZO and BIO+OZO+SF represent
346 minimal abatements relative to the LOQ. Negative abatements were taken into account as zero
347 abatements.
 ACCEPTED MANUSCRIPT

348 Fate of micropollutants during the post-treatments after ozonation

349 Several post-treatments, i.e., sand filtration (SF), moving bed (MB), fixed bed (FB), a fresh granular

350 activated carbon filter (GACfresh/OZO) and a pre-loaded granular activated carbon filter (GACloaded/OZO),

351 were evaluated for their capacity to abate residues of micropollutants still present after ozonation with

352 an ozone dose of 2.7 mg/L ozone (about 0.54 g O3/g DOC). An additional fresh granular activated

PT
353 carbon filter (GACfresh/BIO) was fed with secondary clarifier effluent for a comparison with GACfresh/OZO.

RI
354 Amongst 20 micropollutants (including the indicator substances) frequently detected after ozonation in

355 concentrations >50 ng/L, none showed a significant relative abatement (all <16%) in the three solely

SC
356 biologically active post-treatments, i.e., sand filtration, moving bed and fixed bed (Fig. S8).

357 Similarly, a screening of the selected 550 substances showed that these 3 post-treatments do not

U
significantly impact the number (Fig. S5) and the total concentrations of the compounds detected in the
358
AN
359 ozonated effluent (22.9 µg/L after OZO vs. 24.0 µg/L after SF, 23.5 µg/L after MB and 23.5 µg/L after

360 FB, Fig. S6). The average abatement did not exceed 5% in these biological post-treatments (Fig. S7).
M

361 Even the compounds abated by >50% during the conventional biological treatment (Table S8), i.e.,

362 acesulfame, atenolol, benzotriazole and valsartan, showed a low abatement of <7% in these three post-
D

363 treatments, due to shorter retention times (10-25 min) compared to the conventional biological treatment
TE

364 (18 h).

365 The abatements in two identical fresh activated carbon filters, GACfresh/BIO and GACfresh/OZO, was
EP

366 investigated for four compounds with very good sorption on activated carbon (Group I, see Table S1 for
C

367 definition), two compounds with good sorption (Group II) and two compounds with poor sorption

368 ability (Group III). This grouping is in agreement with the abatement by ozone, and was based on results
AC

369 from sorption to powdered activated carbon, because physical-chemical properties of charged, polar

370 micropollutants like logDow are not sufficient to predict their sorption behavior (Kovalova et al. 2013). A

371 significant difference in the breakthrough of the compounds was observed. Group I compounds (Fig. 2a

372 and b) showed a much later breakthrough compared to the other compounds with weaker sorption

373 behavior (Fig. 2c and d). Surprisingly, benzotriazole from group II was relatively well abated, as

374 discussed later. In GACfresh/OZO, the breakthrough of Groups I, II, III (Fig. 2) and the 10 additional

375 substances telmisartan, valsartan, valsartan acid, primidone, atenolol, atenolol acid, gabapentin,
 ACCEPTED MANUSCRIPT
376 tramadol, lamotrigine, and oxazepam (Fig. S9-S10) usually occurred later than in GACfresh/BIO. For

377 instance, the abatement of group II/III compounds candesartan and sucralose dropped below 90% at

378 about 3’200 BV in GACfresh/BIO but only at about 7’000 BV in GACfresh/OZO. As GACfresh/OZO followed

379 ozonation, the concentrations of the micropollutants were lower and, at the same time, the composition

380 of the matrix is different. It was shown that the breakthrough is independent on micropollutant

PT
381 concentration within certain limits (Zietzschmann et al. 2016). However, during ozonation, effluent

382 organic matter is transformed to smaller compounds, which are generally less aromatic, more

RI
383 hydrophilic and charged, thereby exhibiting a poorer adsorbability on GAC. Measurements showed that

384 small neutral organic substances are further oxidized to charged small organic acids during ozonation

SC
385 (Ramseier and von Gunten 2009, Zietzschmann et al. 2015). As a consequence, micropollutants in

386 ozonated wastewater have fewer competitors for adsorption sites compared to non-ozonated wastewater

387
U
and the breakthrough of micropollutants may occur later in ozonated wastewater.
AN
388 To compare the two GAC filters, the average relative abatement of the indicator substances was

389 evaluated. Because some of the 12 indicator substances are present in too low concentrations after
M

390 ozonation to determine an appropriate abatement, only 6 were selected: 4 compounds from Group I
D

391 (diclofenac, hydrochlorothiazide, metoprolol, venlafaxine), and 2 groups from Group II (benzotriazole
TE

392 and candesartan). This selection criteria is also recommended by the Swiss legislation in case of low

393 inflow concentrations. An average relative abatement of >80% for the 6 selected indicator substances
EP

394 could be kept up to 10’000 BV for GACfresh/BIO and up to 15’000 BV for GACfresh/OZO (Fig. S11). At the

395 two last sampling times for 22’000 and 28’000 BV, the sorption on GACfresh/BIO increased again.
C

396 Currently, this effect cannot be explained and would require more investigations for a clarification.
AC

397 In the screening of the 550 substances, the decrease of total concentration in GACfresh/BIO and

398 GACfresh/OZO was comparable (47% in GACfresh/BIO and 41% in GACfresh/OZO at similar BV in the range of

399 11’000-17’000, Fig. S6). Similarly, the average abatement calculated from the abatement of the same

400 selection of substances was comparable for GACfresh/BIO and GACfresh/OZO (54% and 53%, respectively, at

401 the same BV) (Fig. S7). The performance of GACfresh/BIO can be used to evaluate the effectiveness of

402 GAC filtration. Since this GAC filter has a volume of 77 L and a filling of 34.65 kg GAC, 10’000 BV

403 (where the average abatement of the six indicator substances fell below the limit of 80%) correspond to
 ACCEPTED MANUSCRIPT
404 a carbon usage rate of 45 mg GAC/L. To be comparable to the elimination efficiency of powdered

405 activated carbon, about 10 mg/L activated carbon (corresponding to 45’000 BV) should be reached

406 before breakthrough of 80% of the indicator substances (Boehler et al. 2012b). However, a stepwise

407 renewal of several GAC filter cells running in parallel (with differing elimination efficiencies between

408 100% and below 80%) allows optimal performance with a longer lifetime of the GAC filters (Benstoem

PT
409 et al. 2017). Moreover, it was demonstrated recently that a minimal EBCT is crucial to achieve a good

410 elimination (Wunderlin et al. 2017). An EBCT of 25 minutes (compared to 15 minutes in this study)

RI
411 turned out to be sufficient for a good performance for a 1.18 - 2.36 mm granulation at low DOC.

U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

412

413 Fig. 2. Relative abatements of group I substances (diclofenac, hydrochlorothiazide, metoprolol and venlafaxine) in (a) GACfresh/BIO and (b) GACfresh/OZO;
414 of group II (benzotriazole and candesartan) and group III substances (acesulfame and sucralose) in (c) GACfresh/BIO and (d) GACfresh/OZO as a function of
415 the bed volumes. 28’000 BV correspond to a run time of 10 months.
 ACCEPTED MANUSCRIPT
416 GACloaded/OZO had been pre-loaded for 16’000 BV with the effluent of the secondary clarifier from

417 another WWTP (Boehler et al. 2012a), before being installed at WWTP Neugut after ozonation. During

418 this loading phase of GACloaded/OZO, the abatement of selected micropollutants was measured and can be

419 compared to the abatement in GACfresh/BIO at WWTP Neugut. Micropollutants abatement in GACloaded/OZO

420 (Fig. 3 and Fig. S12, yellow zone) was less efficient than in GACfresh/BIO (Fig. 2 and Fig. S9), even

PT
421 though the EBCT was slightly higher in GACloaded/OZO (18 min for GACloaded/OZO vs. 15 min for

422 GACfresh/BIO). For instance, venlafaxine and sucralose broke through immediately in GACloaded/OZO and

RI
423 their respective abatement at 7’500 BV were around 40% and 30% in GACloaded/OZO, while still at 91%

424 and 68% in GACfresh/BIO, respectively. The wastewater quality (DOC 6.9 mg/L in GACloaded/OZO vs. DOC

SC
425 5.5 mg/L in GACfresh/BIO) and the GAC properties (different material; higher particle size of 1.4-2.36 mm

426 in GACloaded/OZO compared to 0.425-2.36 mm in GACfresh/BIO) likely explain why breakthrough occurred

427 faster in GACloaded/OZO.

U
AN
428 When GACloaded/OZO was operated with ozonated wastewater (>16’000 BV, to the right of the yellow

429 zone in Fig. 3), abatements were not significantly different to the abatements observed during the pre-
M

430 loading phase (yellow zone in Fig. 3). Only diclofenac showed a significant increase in the relative
D

431 abatement in the second phase: its relative abatement was <20% at the end of the first phase at 9’000 BV
TE

432 and 62% when measurements restarted at 23’000 BV. Moreover, the micropollutant’s relative

433 abatements decreased only slightly in GACloaded/OZO over the sampling period between 23’000-50’000
EP

434 BV. Similarly to GACfresh/BIO and GACfresh/OZO, the relative abatement of poorly sorbing compounds such

435 as acesulfame, candesartan, gabapentin, irbesartan, sucralose, and valsartan acid, was <20% at high BVs
C

436 (Fig. 3 and Fig. S12). Conversely, abatement was generally still >60% for diclofenac,
AC

437 hydrochlorothiazide, metoprolol, benzotriazole, methylbenzotriazole and atenolol, even at 50’000 BV.

438 When comparing GACloaded/OZO with GACfresh/OZO at 28’000 BV, the abatement of all substances was

439 similar, with the exception of benzotriazole. This compound was significantly better abated in

440 GACloaded/OZO than in GACfresh/OZO (78% vs. 54%). The results of diclofenac and benzotriazole suggest

441 that, in addition to sorption, a biological degradation may also take place in GACloaded/OZO, in analogy to

442 a previous publication (Reungoat et al. 2012). However, it is not possible to differentiate between

443 sorption and degradation by just measuring changes in concentrations. Fig. S11 shows that the average
 ACCEPTED MANUSCRIPT
444 relative abatement of the 6 selected indicator substances (benzotriazole, candesartan, diclofenac,

445 hydrochlorothiazide, metoprolol, venlafaxine) in GACloaded/OZO was still about 40-50% at 50’000 BV,

446 similar to GACfresh/OZO at 28’000 BV. In the screening of the 550 substances, the average abatement in

447 GACloaded/OZO (19% at 35’000-40’000 BV; see Fig. S7) was less efficient than for the 6 selected indicator

448 substances. The reduction of the total concentration was only 6% in GACloaded/OZO at 35’000-40’000 BV

PT
449 (Fig. S6). This low relative abatement in GACloaded/OZO can be attributed to the very poor abatement of

450 the compounds still present in high concentrations in OZO-EFF (>0.9 µg/L), i.e., acesulfame, sucralose,

RI
451 metformin, iopromide and tris(2-chloropropyl)phosphate, making up 73% of the overall concentration.

452 These compounds are hydrophilic and known to poorly adsorb to activated carbon (Kovalova et al.

SC
453 2013, Margot et al. 2013, Scheurer et al. 2010).

U
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
454
M
D
TE
C EP
AC

455
456 Fig. 3. Relative abatements of (a) group I substances (diclofenac, hydrochlorothiazide, metoprolol
457 and venlafaxine) and (b) group II (benzotriazole and candesartan) and group III substances
458 (acesulfame and sucralose) in GACloaded/OZO. The filter was loaded with the secondary effluent of
459 another WWTP until 16’000 bed volumes (yellow area) and thereafter with ozonated effluent
460 from the WWTP Neugut.
 ACCEPTED MANUSCRIPT
461 Fate of assimilable organic carbon, DOC and oxidation by-products

462 Formation and fate of AOC and fate of DOC

463 Fig. S13 shows that AOC increased during ozonation at the recommended specific ozone dose (0.55 g

464 O3/g DOC) by a factor of 3: the AOC level in BIO-EFF was 90 µg C/L and reached about 280 µg C/L at

PT
465 OZO-EFF. The increase of AOC is associated with the oxidation of dissolved organic matter (DOM) in

466 secondary effluent to form biodegradable and potentially toxic products, such as aldehydes, ketones or

RI
467 carboxylic acids (von Gunten 2003). These classes of compounds are generally well removable during

468 biological post-treatments of ozonated drinking water (Hammes et al. 2006). Among the biological post-

SC
469 treatments studied here, the sand filter performed best in reducing the AOC by 40% (Fig. S14), similar

470 to AOC reduction (14-46%) in other sand filters (Zimmermann et al. 2011). MB and FB reactors only

471
U
slightly reduced AOC (by 17% and 13%, respectively). In FB, an intense growth of snails (Physidae)
AN
472 was observed, preventing the formation of a biofilm, which explains the low performance of this filter.

473 The GAC filters showed the highest abatement of AOC (72% in GACfresh/OZO at 15’000 BV and 57% in
M

474 GACloaded/OZO at 35’000 BV) due to additional sorption capability.

D

475 Ozonation alone induced a relative abatement of DOC of about 10% (Fig. S15), which is relatively high
TE

476 in comparison to other studies where the DOC concentration almost did not change (Haslinger et al.

477 submitted, Hollender et al. 2009). The DOC concentration after ozonation was poorly abated in in MB
EP

478 and FB (6% and 5%, respectively). In contrast, the sand filter showed a better relative abatement of

479 DOC (20%), due to biodegradation of degradable compounds and the deposition of colloidal organic
C

480 substances on the sand filter material. The GAC filters performed similar to sand filtration with a
AC

481 relative DOC abatement of 15-20% on GACfresh/OZO or GACloaded/OZO at 20’000-50’000 BV. The loading

482 of DOC at the beginning (up to 15’000 BV) onto GACfresh/bio with biologically treated wastewater

483 occurred slightly faster than onto GACfresh/OZO with the ozonated effluent (Fig. S15-S16). The same

484 behavior was observed by Haslinger et al. (submitted).

485 Similar to DOC, the removal of total suspended solids (TSS) was less efficient in MB or FB (4% or

486 17%, respectively) than in the filter systems, where 75% TSS removal was achieved in SF and 70-80%
 ACCEPTED MANUSCRIPT
487 in the GAC filters (Fig. S17). A filter cloth installed after the MB could enhance the overall TSS

488 elimination to 50%. TSS was relatively low at around 2 mg/L in the influent and effluent of ozonation.

489 Formation and fate of bromate

490 The concentration of bromide in the influent to the ozonation reactor (n=30) was on average 60 µg/L

PT
491 ranging 35-85 µg/L (n=26), but a few exceptionally high values (up to 150 µg/L) were measured (Table

492 S10). These values correspond to common bromide concentrations in Swiss municipal WWTPs without

RI
493 discharges from waste incinerators, chemical industries or landfills (Schindler Wildhaber et al. 2015).

494 Bromate, an ozonation product from the oxidation of bromide, was also detected in WWTP-INF and

SC
495 BIO-EFF, mostly below 1 µg/L (n=20), with a maximum concentration of 17 µg/L, concurring with a

496 relatively high bromide concentration (85 µg/L). The presence of bromate before ozonation may be

497
U
attributed to a pre-oxidation in one of the industrial sites connected to the WWTP.
AN
498 During ozonation, bromide was oxidized to bromate with increasing yields for increasing specific ozone

499 doses, reaching 5.7 µg/L at 0.97 g O3/g DOC with a bromide level of 60 µg/L (Fig. S18). The bromate
M

500 yield was calculated as the ratio between the produced bromate and the bromide concentration in the
D

501 influent (Fig. 4). For specific ozone doses ≤0.45 g O3/g DOC, the bromate yield was low (<1%, w/w).
TE

502 In this range, ozone and •OH exposures are very low, because ozone is rapidly depleted due to the

503 multiple reactions of ozone with DOM (Chon et al. 2015, Soltermann et al. 2017). For specific ozone
EP

504 doses >0.45 g O3/g DOC, the lifetime of ozone increases steadily with an increasing bromate yield

505 reaching 14% for 1.07 g O3/g DOC. For the recommended specific ozone dose of 0.55 g O3/g DOC, the
C

506 bromate yield is approximately 3%. At this ozone dose, the bromate formation corresponds to 1.8 µg/L
AC

507 for the average concentration of bromide (60 µg/L), and in worse case 4.5 µg/L at the rare maximum

508 concentration of bromide (150 µg/L). As expected, Fig. S19 shows no significant changes of bromide or

509 bromate concentration in all post-treatments. Since bromate was mostly below 1 µg/L in BIO-EFF, its

510 concentration in the WWTP effluent would range normally between 1 and 3.5 µg/L at bromide levels

511 below 85 µg/L. Taking into account that, after discharge in the river, wastewater effluent is diluted by an

512 average factor of 20 at this site, the WWTP input on bromate concentration in the receiving water is low

513 (normally below 0.18 µg/L, with a maximal level of 0.28 µg/L at the highest bromide concentration),
 ACCEPTED MANUSCRIPT
514 well below the environmental quality standard (50 µg/L) (Ecotox Center 2017), the usual drinking water

515 standards (10 µg/L) (Commission Directive 2003, U.S. EPA 2006, WHO 2011), and even below the

516 most stringent guidelines (1 µg/L) applied for instance in the Netherlands (Knol et al. 2015). In the

517 current study, once a bromate concentration of 17 µg/L (Table S10) was observed in WWTP-INF, which

518 would lead to a bromate concentration as high as 21.5 µg/L in the effluent. In this case, the sources of

PT
519 bromate should be identified and measures should be taken to minimize the bromate discharge.

RI
U SC
AN
M
D
TE
EP

520
521 Fig. 4. Bromate yield (w/w) in the full-scale ozonation for a range of specific ozone doses from
522 0.32-1.07 g O3/g DOC (n=30).
C

523 Formation and fate of nitrosamines

AC

524 NDMA was regularly detected in relatively high concentrations after the biological treatment (BIO-

525 EFF), with a maximum concentration of 190 ng/L. This is in accordance with previous measurements in

526 20 WWTPs in Switzerland where after the biological treatment NDMA concentrations were found from

527 <LOQ to 188 ng/L (Krauss et al. 2009). In the US, NDMA concentrations in non-chlorinated effluents

528 were found up to 380 ng/L (Pehlivanoglu-Mantas et al. 2006, Plumlee and Reinhard 2007, Sedlak et al.

529 2005). A large variation was also observed from day to day and among differing WWTPs. At WWTP

530 Neugut, the maximum NDMA formation during ozonation for specific ozone doses in the range of 0.54-
 ACCEPTED MANUSCRIPT
531 0.97 g O3/g DOC (Fig. 5) was 30 ng/L. No correlation was observed between the formation of NDMA

532 and the specific ozone dose, in agreement to other studies (Hollender et al. 2009, Schindler Wildhaber et

533 al. 2015, Zimmermann et al. 2011). Thus, the secondary clarifier effluent composition with likely

534 changing concentrations of differing NDMA precursors (not determined here) and varying NDMA

535 formation rates affects the extent of NDMA more than the specific ozone dose. In sand filtration,

PT
536 NDMA was significantly abated always reaching values below 50 ng/L in the WWTP effluent (Fig. 5).

537 Therefore, the expected concentration in the receiving waters after a 20-fold dilution is well below the

RI
538 WHO drinking water guideline of 100 ng/L and also below the tentative drinking water value of 10 ng/L

539 in several countries like Germany (Selin 2011).

U SC
AN
M
D
TE
C EP
AC

540
541 Fig. 5. Concentrations of NDMA in BIO-EFF, OZO-EFF and SF-EFF for various specific ozone
542 doses ranging between 0.54-0.97 g O3/g DOC. *: 72-h flow proportional samples; **: 48-h flow
543 proportional samples; ***: 24-h flow proportional samples.
544

545 NDMA was abated in all of the investigated post-treatments (Fig. S20). Among the solely biological

546 post-treatments, SF showed the best relative degradation of NDMA by 66%, while MB and FB showed

547 slightly lower relative eliminations (48% and 41%, respectively). The relatively fresh filter GACfresh/OZO
 ACCEPTED MANUSCRIPT
548 (12’500-16’500 BV) showed the highest abatement of 83%, while GACloaded/OZO at higher BV (35’000-

549 40’000 BV) exhibited a similar efficiency as SF (with 56% removal of NDMA).

550 Fate of transformation products

551 Formation of transformation products during ozonation

PT
552 Sixty-three ozonation transformation products from a total of 19 parents were screened for, based on

553 evidence of formation in literature (Table S6). Eighteen of them could be quantified because standards

RI
554 were available. However, ten out of these (3-[(4-chlorobenzoyl)amino] propanoic acid, 1-[(4-

SC
555 chlorophenyl) phenylmethyl] piperazine, 4-chlorobenzophenone, N-desmethyl-clarithromycin, DCPMU,

556 azacyclonol, gabapentin acid, ranitidine S-oxide, N-desmethyl-tramadol and N,N-didesmethyl-tramadol)

U
557 were detected in none of the ozonated wastewater samples. 4’-/5-hydroxydiclofenac was detected in
AN
558 high concentrations in most of the wastewater samples, however, its concentration dramatically

559 decreased after ozonation, independently of the specific ozone dose, and was therefore not evaluated
M

560 further. The following discussion will focus on seven OTPs frequently detected in wastewater samples.

561 Some of these OTPs are also known to be human metabolites, biological transformation products or
D

562 used as active ingredients, and were therefore already detected in the wastewater after the conventional
TE

563 biological treatment (Table S5). For instance, chlorothiazide was detected in all BIO-EFF samples

564 (n=12) at concentrations ranging 45-67 ng/L during the evaluation of the 4 specific ozone doses.
EP

565 Tramadol N-oxide (n=5) , clarithromycin N-oxide (n=3) and cetirizine N-oxide (n=1) were detected at

566 least once in samples from biological treatment, at a maximum concentration of 35 ng/L.
C

567 The formation of the seven OTPs and their molar yields decreased with increasing specific ozone doses
AC

568 (Fig. S21 and Table S11). For instance, in the case of chlorothiazide, the concentration of its parent

569 compound, hydrochlorothiazide, was relatively constant in BIO-EFF over the sampling period (ca. 1000

570 ng/L), but chlorothiazide concentrations dropped from 421-446 ng/L at 0.35 g O3/g DOC to 163-308

571 ng/L at 0.97 g O3/g DOC, respectively, corresponding to a decrease of the molar yield from 40% to

572 21%. At the specific ozone dose of 0.54 gO3/gDOC, only clarithromycin N-oxide (62-74 ng/L),

573 chlorothiazide (364-385 ng/L), fexofenadine N-oxide (111-171 ng/L), tramadol N-oxide (76-82 ng/L)

574 and venlafaxine N-oxide (27-30 ng/L) were >LOQ in OZO-EFF. Highest molar yields of the OTPs were
 ACCEPTED MANUSCRIPT
575 observed for chlorothiazide (36%), clarithromycin N-oxide (34%), and fexofenadine N-oxide (30%) at

576 this specific ozone dose. Amisulpride N-oxide and cetirizine N-oxide were detected only at 0.35 g O3/g

577 DOC (n=3), but their yields did not exceed 4% and 3%, respectively. The selected OTPs were formed

578 rapidly during ozonation due to the high ozone-reactivity of the parent compounds (kO3>104 M-1 s-1) as

579 recently determined in batch experiments with ozone alone for chlorothiazide, fexofenadine N-oxide and

PT
580 cetirizine N-oxide (Borowska et al. 2016), but were subsequently further transformed mostly by attack

581 of hydroxyl radical.

RI
582 For the 45 OTPs without available standard, normalized peak areas were determined in the samples (Fig.

583 S22). Out of these potential OTPs, the exact masses of only 12 were detected after ozonation: four OTPs

SC
584 from hydrochlorothiazide (HCT_TP1-4, Table S6), two from carbamazepine (BQM and BaQD), two

585 from phenazone (PHE_TP2 and TP6), two from sulfamethoxazole (SMX_TP1 and TP3), one from

586
U
atenolol (ATE_TP1), and one from lamotrigine (LAM_TP2). These results are in agreement with the
AN
587 studies listed in Table S6, where these OTPs were found to be predominantly formed during the

588 ozonation of parent compounds.

M

589 Fate of the transformation products during post-treatments

D
TE

590 Among the seven quantified OTPs, only five OTPs (chlorothiazide, clarithromycin N-oxide, tramadol N-

591 oxide, venlafaxine N-oxide and fexofenadine-N-oxide) were measured regularly in concentrations well
EP

592 above the LOQ and could be evaluated in the post-treatments. SF, MB and FB were all inefficient to

593 biodegrade these OTPs (Fig. S23). This shows that the investigated OTPs are very stable compounds,
C

594 not being degradable in the selected post-treatments.

AC

595 In contrast to the three biological post-treatments, GACfresh/OZO was efficient to remove the selected

596 OTPs due to sorption (Fig. 6 and Fig. S24). Up to about 8’000 BV, the abatement was ≥80% for N-

597 oxides and even >96% for chlorothiazide. Thereafter, the abatement of the N-oxides started to decrease,

598 but stayed above 80% until 17’000 BV for chlorothiazide. At 23’000 BV, the abatement of N-oxides

599 was <40%, while chlorothiazide abatement was still higher (57%). In GACloaded/OZO, only chlorothiazide

600 was efficiently abated by 30-40% at 50’000 BV (Fig. S23b). In general, more polar compounds are

601 produced in ozonation, which are expected to be less efficiently sorbed than their parent compounds,
 ACCEPTED MANUSCRIPT
602 therefore exhibiting an earlier breakthrough. For instance, the sorption of venlafaxine N-oxide and

603 tramadol N-oxide on GACfresh/OZO dropped below 80% at 8’000 BV, while for the parent compounds

604 venlafaxine and tramadol, this occurred only after 14’000 BV (Fig. 6). The relative abatement of

605 chlorothiazide was also smaller than for hydrochlorothiazide, but only at higher BVs. However, the

606 abatement of fexofenadine was similar to fexofenadine N-oxide (the biggest molecule of the three N-

PT
607 oxides, mostly uncharged), where the difference in polarity between the parent and the OTP is probably

608 less pronounced.

RI
609 OTPs without available standards also showed mostly no significant changes of normalized peak areas

610 in SF, MB and FB (Fig. S22). However, significant peak reductions were observed in GACfresh/OZO (at

SC
611 14’000 BV) and rarely in GACloaded/OZO (at 37’000 BV), which may be explained by the respective

612 residual sorption capacity. Transformation products of carbamazepine, BQM and BaQD, showed a

613
U
particular behavior in SF and GACloaded/OZO: The BQM peak area was reduced concomitantly to a BaQD
AN
614 peak increase. The other biological systems, MB and FB, showed no change. These results are in

615 agreement with previous observations (Hübner et al. 2014), where BQM formation was found as the
M

616 major OTP of carbamazepine, and a successive conversion of BQM to BaQM (not detected in this
D

617 study) and of BaQM to BaQD was observed in the subsequent biological post-treatments.
TE
C EP
AC
 ACCEPTED MANUSCRIPT

PT
RI
U SC
AN
M
D
TE
C EP
AC

618

619 Fig. 6. Comparison of relative abatement of parent compounds (solid) and their respective primary OTP (empty) in GACfresh/OZO (triangle) and
620 GACloaded/OZO (square) for (a) hydrochlorothiazide/chlorothiazide, (b) tramadol/tramadol N-oxide, (c) venlafaxine/venlafaxine N-oxide, and (d)
621 fexofenadine/ fexofenadine N-oxide. Tramadol*: sum of tramadol and O-desmethylvenlafaxine (see Text S2).
 ACCEPTED MANUSCRIPT
622 Cost of new ozone treatment

623 The cost estimation of the installation and operation of ozonation at WWTP Neugut is presented in

624 detail in SI Text S6. To summarize, investments accounted to 3.27 Mio CHF, and operation costs are

625 240 CHF/d or 0.013 CHF/m3 (180 CHF/d for ozone and electricity and 60 CHF/d for manpower).

626 Overall, the total cost accounts to 0.06 CHF/m3, corresponding to an increase of 10% for wastewater

PT
627 treatment at WWTP Neugut, what is at the lower end of general estimations (Abegglen and Siegrist

628 2012). According to the new Swiss Water Protection Act, the investment costs will be covered to 75%

RI
629 by a fund established from payments of 9 CHF/y by each Swiss inhabitant who is connected to a WWTP

630 without installation of advanced treatment (Eggen et al. 2014).

U SC
AN
M
D
TE
C EP
AC
 ACCEPTED MANUSCRIPT
631 CONCLUSIONS

632 • Ozonation of the effluent of the secondary clarifier in WWTP Neugut resulted in a significant

633 reduction of the micropollutant load in the WWTP effluent. At the recommended ozone dose

634 (0.55 g O3/g DOC), indicator substances were abated by >80% even at potential high nitrite

635 concentrations of 0.2 mg NO2--N/L. For a set of about 200 micropollutants detected in the

PT
636 influent of the WWTP, an average abatement of >79% was observed over the whole treatment

637 chain. Nitrite could be compensated directly by using a feed-back control for ozone dosing by

RI
638 the delta UV254nm signal from in- and outlet of the ozone reactor (Wittmer et al. 2015). In this

639 case, a specific ozone dose of 0.40 g O3/g DOC (1.5-2.5 mg O3/L) would be sufficient for the

SC
640 required micropollutant abatement of 80%.

641 • For the implementation of ozonation, the formation of the oxidation by-products bromate and

642
U
NDMA needed to be evaluated and were found to be formed also at WWTP Neugut. NDMA
AN
643 was significantly reduced in all post-treatments. As bromide was mostly in a concentration

range common for Swiss municipal wastewater, the bromate level was typically < 3.5 ug/L in
M

644

645 the WWTP effluent. The concentrations expected after dilution of discharged water were clearly
D

646 below drinking water standards (10 µg/L for bromate, 10 ng/L for NDMA).
TE

647 • While the micropollutants were not significantly abated in the biological post-treatment (SF,

648 MB, FB) as expected, GAC filtration was efficient to remove the majority of micropollutants
EP

649 due to sorption, even after 30’000 bed volumes.

650 • GACfresh/OZO (operated with wastewater treated biologically and subsequently by ozonation)
C

651 showed a later breakthrough of DOC and micropollutants than GACfresh/BIO (operated with
AC

652 biologically treated wastewater), potentially due to altered organic matter, being less

653 competitive for sorption sites after ozonation.

654 • GACloaded/OZO was efficient to abate certain substances even up to 50’000 bed volumes (about 2

655 years of running time). It cannot be assessed so far whether this is due to a residual sorption

656 capacity of the activated carbon or due to biodegradation. GAC filters operated at high bed

657 volumes should be further investigated as they represent a promising alternative to sand filters

658 for an improved quality of WWTP effluents.

 ACCEPTED MANUSCRIPT
659 • The investigated OTPs were not abated in the biological post-treatments, except one OTP of

660 carbamazepine, which was degraded to another OTP. The detected OTPs appeared to be very

661 stable compounds, and biodegradable compounds seem to be more difficult to be detected. Only

662 the GAC filters were showing abatement of OTPs due to additional sorption capacities, even at

663 30’000 bed volumes. However, as demonstrated by the abatement of AOC in the biological

PT
664 post-treatments, degradable compounds are being removed, complying with the task of the post-

665 treatments. Overall, the main goal of the biological post-treatment is the elimination of possible

RI
666 negative ecotoxicological effects generated during ozonation.

667 • For the evaluation of ozonation and the post-treatments, the water quality needs to be assessed

SC
668 with bioassays. Therefore, various ecotoxicological endpoints, using living cells, organisms and

669 communities, were investigated. The outcomes of these tests will be presented in a follow-up

670 study (Kienle et al. in preparation).

U
AN
671 Acknowledgements
M

672 The authors would like to thank the EU within FP7 (Demeau project, Grant Agreement no. 308339), the

673 Federal Office for the Environment FOEN (ReTreat and UV projects) and SCIEX- Scientific Exchange
D

674 Program NMS (Microzo project, Project code 12.333) for fundings. Project partners (Chemviron,
TE

675 Ensola, Mecana, Ozonia and Wabag) are acknowledged for their participation in the ReTreat project.

676 Max Schachtler and co-workers (WWTP Neugut), Michael Thomann (Holinger AG) and Michel
EP

677 Blunschi (Ingenieurbüro Gujer AG) are acknowledged for their collaboration. The authors would like to

678 thank particularly Juliane Hollender, Heinz Singer, Anita Wittmer (Eawag) and Cornelia Kienle (Swiss
C

679 Centre for Applied Ecotoxicology Eawag-EPFL) for their contributions to the fruitful discussions within
AC

680 this project.

681
682 Supporting information

683 Additional data associated with this article can be found in the Supplementary Material (Text S1-S5,

684 Fig. S1-S24 and Table S1-S11).

 ACCEPTED MANUSCRIPT
685 REFERENCES

686 Abegglen, C. and Siegrist, H. (2012) Mikroverunreinigungen aus kommunalem Abwasser. Verfahren zur

687 weitergehenden Elimination auf Kläranlagen. Bundesamt für Umwelt, Bern, Umwelt-Wissen Nr. 1214: 210

688 S.

689 Benstoem, F., Nahrstedt, A., Boehler, M., Knopp, G., Montag, D., Siegrist, H. and Pinnekamp, J. (2017)

PT
690 Performance of granular activated carbon to remove micropollutants from municipal wastewater—A meta-

analysis of pilot- and large-scale studies. Chemosphere 185, 105-118.

RI
691

692 Boehler, M., Wittmer, A., Heisele, A., Wohlhausser, A., Salhi, L., von Gunten, U., Siegrist, H., McArdell,

SC
693 C., Longrée, P. and Beck, B. (2012a) Ergänzende Untersuchungen zur Elimination von

694 Mikroverunreinigungen auf der Ara Neugut.

U
AN
695 Boehler, M., Zwickenpflug, B., Hollender, J., Ternes, T., Joss, A. and Siegrist, H. (2012b) Removal of

696 micropollutants in municipal wastewater treatment plants by powder-activated carbon. Water Sci. Technol.
M

697 66(10), 2115-2121.

D

698 Bollmann, A.F., Seitz, W., Prasse, C., Lucke, T., Schulz, W. and Ternes, T. (2016) Occurrence and fate of
TE

699 amisulpride, sulpiride, and lamotrigine in municipal wastewater treatment plants with biological treatment

700 and ozonation. J. Hazard. Mater. 320, 204-215.

EP

701 Buxton, G.V. and Elliot, A.J. (1986) Rate-Constant for Reaction of Hydroxyl Radicals with Bicarbonate

702 Ions. Radiat. Phys. Chem. 27(3), 241-243.

C
AC

703 Buxton, G.V., Greenstock, C.L., Helman, W.P. and Ross, A.B. (1988) Critical Review of rate constants for

704 reactions of hydrated electrons, hydrogen atoms and hydroxyl radicals (⋅OH/⋅O−) in Aqueous Solution. J.

705 Phys. Chem. Ref. Data 17(2), 513-886.

706 California Department of Public Health (2010) Drinking Water Notification Levels and Response Levels: An

707 Overview.

708 http://www.cdph.ca.gov/certlic/drinkingwater/Documents/Notificationlevels/notificationlevels.pdf
 ACCEPTED MANUSCRIPT
709 Chon, K., Salhi, E. and von Gunten, U. (2015) Combination of UV absorbance and electron donating

710 capacity to assess degradation of micropollutants and formation of bromate during ozonation of wastewater

711 effluents. Water Res. 81, 388-397.

712 Commission Directive (2003) Official Journal of the European Union, Establishing the List, Concentration

713 Limits and Labeling Requirements for the Constituents of Natural Mineral Waters and the Conditions for

PT
714 Using Ozone-Enriched Air for the Treatment of Natural Mineral Waters and Spring Waters.

RI
715 Ecotox Center (2017) Environmental Quality Standard (EQS) - Vorschlag des Oekotoxzentrums für Bromat.

716 http://www.ecotoxcentre.ch/expert-service/quality-standards/proposals-for-acute-and-chronic-quality-

SC
717 standards/ (accessed 1 August 2017)

U
718 Eggen, R.I.L., Hollender, J., Joss, A., Scharer, M. and Stamm, C. (2014) Reducing the Discharge of
AN
719 Micropollutants in the Aquatic Environment: The Benefits of Upgrading Wastewater Treatment Plants.

720 Environ. Sci. Technol. 48(14), 7683-7689.

M

721 European Chemical Agency (2017) https://echa.europa.eu/chemicals-in-our-life/how-european-legislation-

D

722 on-chemicals-improves-our-lives/free-access-to-data (accessed 1 August 2017)

TE

723 FOEN (2017) Gewässerqualität: Revision der Gewässerschutzverordnung.

724 https://www.bafu.admin.ch/bafu/fr/home/themes/formation/communiques.msg-id-59323.html (accessed 1

EP

725 August 2017)

C

726 Gros, M., Petrovic, M., Ginebreda, A. and Barcelo, D. (2010) Removal of pharmaceuticals during
AC

727 wastewater treatment and environmental risk assessment using hazard indexes. Environ. Int. 36(1), 15-26.

728 Haag, W.R. and Hoigné, J. (1983) Ozonation of bromide-containing waters: kinetics of formation of

729 hypobromous acid and bromate. Environ. Sci. Technol. 17(5), 261-267.

730 Hammes, F., Salhi, E., Köster, O., Kaiser, H.-P., Egli, T. and von Gunten, U. (2006) Mechanistic and kinetic

731 evaluation of organic disinfection by-product and assimilable organic carbon (AOC) formation during the

732 ozonation of drinking water. Water Res. 40(12), 2275-2286.

 ACCEPTED MANUSCRIPT
733 Haslinger, J., Schaar, H., Saracevic, E., Walder, C. and Kreuzinger, N. (submitted) Impact of ozonation on

734 adsorptive removal of organic micropollutants in granular activated carbon filters. J. Environ. Chem. Eng.

735 Herzog, B., Lemmer, H., Huber, B., Horn, H. and Müller, E. (2014) Xenobiotic benzotriazoles—

736 biodegradation under meso- and oligotrophic conditions as well as denitrifying, sulfate-reducing, and

737 anaerobic conditions. Environ. Sci. Pollut. 21(4), 2795-2804.

PT
738 Hoigné, J. and Bader, H. (1975) Ozonation of water - Role of hydroxyl radicals as oxidizing intermediates.

RI
739 Science 190(4216), 782-784.

SC
740 Hollender, J., Zimmermann, S.G., Koepke, S., Krauss, M., McArdell, C.S., Ort, C., Singer, H., von Gunten,

741 U. and Siegrist, H. (2009) Elimination of Organic Micropollutants in a Municipal Wastewater Treatment

U
742 Plant Upgraded with a Full-Scale Post-Ozonation Followed by Sand Filtration. Environ. Sci. Technol.
AN
743 43(20), 7862-7869.

744 Hübner, U., Seiwert, B., Reemtsma, T. and Jekel, M. (2014) Ozonation products of carbamazepine and their
M

745 removal from secondary effluents by soil aquifer treatment – Indications from column experiments. Water
D

746 Res. 49, 34-43.

TE

747 Hübner, U., von Gunten, U. and Jekel, M. (2015) Evaluation of the persistence of transformation products

748 from ozonation of trace organic compounds – A critical review. Water Res. 68, 150-170.
EP

749 Jeon, J., Kurth, D. and Hollender, J. (2013) Biotransformation pathways of biocides and pharmaceuticals in
C

750 freshwater crustaceans based on structure elucidation of metabolites using high resolution mass
AC

751 spectrometry. Chem. Res. Toxicol. 26(3), 313-324.

752 Kern, S., Baumgartner, R., Helbling, D.E., Hollender, J., Singer, H., Loos, M.J., Schwarzenbach, R.P. and

753 Fenner, K. (2010) A tiered procedure for assessing the formation of biotransformation products of

754 pharmaceuticals and biocides during activated sludge treatment. J. Environ. Monitor. 12(11), 2100-2111.
 ACCEPTED MANUSCRIPT
755 Kern, S., Fenner, K., Singer, H.P., Schwarzenbach, R.P. and Hollender, J. (2009) Identification of

756 transformation products of organic contaminants in natural waters by computer-aided prediction and high-

757 resolution mass spectrometry. Environ. Sci. Technol. 43(18), 7039-7046.

758 Kienle, C., Boehler, M., Bourgin, M., Fleiner, J., Simon, E., Teichler, R., von Gunten, U., Siegrist, H. and

759 McArdell, C.S. (in preparation) Evaluation of a full-scale wastewater treatment plant upgraded with

PT
760 ozonation and biological post-treatments: Ecotoxicological evaluation. Water Res.

RI
761 Knol, A.H., Lekkerkerker-Teunissen, K., Houtman, C.J., Scheideler, J., Ried, A. and van Dijk, J.C. (2015)

762 Conversion of organic micropollutants with limited bromate formation during the Peroxone process in

SC
763 drinking water treatment. Drink. Water Eng. Sci. 8(2), 25-34.

U
764 Kosaka, K., Asami, M., Konno, Y., Oya, M. and Kunikane, S. (2009) Identification of Antiyellowing Agents
AN
765 as Precursors of N-Nitrosodimethylamine Production on Ozonation from Sewage Treatment Plant Influent.

766 Environ. Sci. Technol. 43(14), 5236-5241.

M

767 Kovalova, L., Siegrist, H., Singer, H., Wittmer, A. and McArdell, C.S. (2012) Hospital Wastewater
D

768 Treatment by Membrane Bioreactor: Performance and Efficiency for Organic Micropollutant Elimination.
TE

769 Environ. Sci. Technol. 46(3), 1536-1545.

770 Kovalova, L., Siegrist, H., von Gunten, U., Eugster, J., Hagenbuch, M., Wittmer, A., Moser, R. and
EP

771 McArdell, C.S. (2013) Elimination of micropollutants during post-treatment of hospital wastewater with

772 powdered activated carbon, ozone, and UV. Environ. Sci. Technol. 47(14), 7899-7908.
C
AC

773 Krasner, S.W., Mitch, W.A., McCurry, D.L., Hanigan, D. and Westerhoff, P. (2013) Formation, precursors,

774 control, and occurrence of nitrosamines in drinking water: A review. Water Res. 47(13), 4433-4450.

775 Krauss, M. and Hollender, J. (2008) Analysis of Nitrosamines in Wastewater:  Exploring the Trace Level

776 Quantification Capabilities of a Hybrid Linear Ion Trap/Orbitrap Mass Spectrometer. Anal. Chem. 80(3),

777 834-842.
 ACCEPTED MANUSCRIPT
778 Krauss, M., Longree, P., Dorusch, F., Ort, C. and Hollender, J. (2009) Occurrence and removal of N-

779 nitrosamines in wastewater treatment plants. Water Res. 43(17), 4381-4391.

780 Kurokawa, Y., Maekawa, A., Takahashi, M. and Hayashi, Y. (1990) Toxicity and carcinogenicity of

781 potassium bromate--a new renal carcinogen. Environ. Health Perspect. 87, 309-335.

PT
782 Lee, M., Blum, L.C., Schmid, E., Fenner, K. and von Gunten, U. (2017) A computer-based prediction

783 platform for the reaction of ozone with organic compounds in aqueous solution: kinetics and mechanisms.

RI
784 Environ. Sci. Process. Impacts 19(3), 465-476.

SC
785 Lee, Y., Gerrity, D., Lee, M., Bogeat, A.E., Salhi, E., Gamage, S., Trenholm, R.A., Wert, E.C., Snyder, S.A.

786 and von Gunten, U. (2013) Prediction of Micropollutant Elimination during Ozonation of Municipal

U
787 Wastewater Effluents: Use of Kinetic and Water Specific Information. Environ. Sci. Technol. 47(11), 5872-
AN
788 5881.

789 Lee, Y. and von Gunten, U. (2016) Advances in predicting organic contaminant abatement during ozonation
M

790 of municipal wastewater effluent: reaction kinetics, transformation products, and changes of biological
D

791 effects. Environ. Sci-Wat. Res. 2(3), 421-442.

TE

792 Lim, S., Lee, W., Na, S., Shin, J. and Lee, Y. (2016) N-nitrosodimethylamine (NDMA) formation during

793 ozonation of N,N-dimethylhydrazine compounds.: Reaction kinetics, mechanisms, and implications for
EP

794 NDMA formation control. Water Res. 105, 119-128.

C

795 Liu, Y.-S., Ying, G.-G., Shareef, A. and Kookana, R.S. (2011) Biodegradation of three selected
AC

796 benzotriazoles under aerobic and anaerobic conditions. Water Res. 45(16), 5005-5014.

797 Logar, I., Brouwer, R., Maurer, M. and Ort, C. (2014) Cost-Benefit Analysis of the Swiss National Policy on

798 Reducing Micropollutants in Treated Wastewater. Environ. Sci. Technol. 48(21), 12500-12508.

799 Loos, R., Carvalho, R., António, D.C., Comero, S., Locoro, G., Tavazzi, S., Paracchini, B., Ghiani, M.,

800 Lettieri, T., Blaha, L., Jarosova, B., Voorspoels, S., Servaes, K., Haglund, P., Fick, J., Lindberg, R.H.,
 ACCEPTED MANUSCRIPT
801 Schwesig, D. and Gawlik, B.M. (2013) EU-wide monitoring survey on emerging polar organic contaminants

802 in wastewater treatment plant effluents. Water Res. 47(17), 6475-6487.

803 Lopez, B., Ollivier, P., Togola, A., Baran, N. and Ghestem, J.P. (2015) Screening of French groundwater for

804 regulated and emerging contaminants. Sci. Total Environ. 518-519, 562-573.

PT
805 Margot, J., Kienle, C., Magnet, A., Weil, M., Rossi, L., de Alencastro, L.F., Abegglen, C., Thonney, D.,

806 Chèvre, N., Schärer, M. and Barry, D.A. (2013) Treatment of micropollutants in municipal wastewater:

RI
807 Ozone or powdered activated carbon? Sci. Total Environ. 461–462, 480-498.

SC
808 Massachusetts Department of Environmental Protection (2004) Current Regulatory Limit: N-

809 Nitrosodimethylamine. www.mass.gov/eea/agencies/massdep/water/drinking/standards/n-

U
810 nitrosodimethylamine-ndma.html
AN
811 Mazioti, A.A., Stasinakis, A.S., Gatidou, G., Thomaidis, N.S. and Andersen, H.R. (2015) Sorption and

812 biodegradation of selected benzotriazoles and hydroxybenzothiazole in activated sludge and estimation of
M

813 their fate during wastewater treatment. Chemosphere 131, 117-123.

D

814 Pehlivanoglu-Mantas, E., Hawley, E.L., Deeb, R.A. and Sedlak, D.L. (2006) Formation of
TE

815 nitrosodimethylamine (NDMA) during chlorine disinfection of wastewater effluents prior to use in irrigation

816 systems. Water Res. 40(2), 341-347.

EP

817 Petrie, B., Barden, R. and Kasprzyk-Hordern, B. (2014) A review on emerging contaminants in wastewaters
C

818 and the environment: Current knowledge, understudied areas and recommendations for future monitoring.
AC

819 Water Res. 72, 3-27.

820 Planas, C., Palacios, Ó., Ventura, F., Rivera, J. and Caixach, J. (2008) Analysis of nitrosamines in water by

821 automated SPE and isotope dilution GC/HRMS: Occurrence in the different steps of a drinking water

822 treatment plant, and in chlorinated samples from a reservoir and a sewage treatment plant effluent. Talanta

823 76(4), 906-913.

 ACCEPTED MANUSCRIPT
824 Plumlee, M.H. and Reinhard, M. (2007) Photochemical attenuation of N-nitrosodimethylamine (NDMA) and

825 other Nitrosamines in surface water. Environ. Sci. Technol. 41(17), 6170-6176.

826 Prasse, C., Schlusener, M.P., Schulz, R. and Ternes, T.A. (2010) Antiviral Drugs in Wastewater and Surface

827 Waters: A New Pharmaceutical Class of Environmental Relevance? Environ. Sci. Technol. 44(5), 1728-

828 1735.

PT
829 Ramseier, M.K. and von Gunten, U. (2009) Mechanisms of phenol ozonation-Kinetics of formation of

RI
830 primary and secondary reaction products. Ozone Sci. Eng. 31(3), 201-215.

SC
831 Reungoat, J., Escher, B.I., Macova, M., Argaud, F.X., Gernjak, W. and Keller, J. (2012) Ozonation and

832 biological activated carbon filtration of wastewater treatment plant effluents. Water Res. 46(3), 863-872.

833
U
Reungoat, J., Macova, M., Escher, B.I., Carswell, S., Mueller, J.F. and Keller, J. (2010) Removal of
AN
834 micropollutants and reduction of biological activity in a full scale reclamation plant using ozonation and

835 activated carbon filtration. Water Res. 44(2), 625-637.

M

836 Richardson, S.D. and Kimura, S.Y. (2016) Water Analysis: Emerging Contaminants and Current Issues.
D

837 Anal. Chem. 88(1), 546-582.

TE

838 Scheurer, M., Storck, F.R., Brauch, H.-J. and Lange, F.T. (2010) Performance of conventional multi-barrier
EP

839 drinking water treatment plants for the removal of four artificial sweeteners. Water Res. 44(12), 3573-3584.

840 Schindler Wildhaber, Y., Mestankova, H., Scharer, M., Schirmer, K., Salhi, E. and von Gunten, U. (2015)
C

841 Novel test procedure to evaluate the treatability of wastewater with ozone. Water Res. 75, 324-335.
AC

842 Schmidt, C.K. and Brauch, H.J. (2008) N,N-dimethosulfamide as precursor for N-nitrosodimethylamine

843 (NDMA) formation upon ozonation and its fate during drinking water treatment. Environ. Sci. Technol.

844 42(17), 6340-6346.

845 Sedlak, D.L., Deeb, R.A., Hawley, E.L., Mitch, W.A., Durbin, T.D., Mowbray, S. and Carr, S. (2005)

846 Sources and fate of nitrosodimethylamine and its precursors in municipal wastewater treatment plants. Water

847 Environ. Res. 77(1), 32-39.

 ACCEPTED MANUSCRIPT
848 Selin, N.E. (2011) Environmental guidelines and regulations for nitramines: a policy summary.

849 http://www.tcmda.com/Global/Aminrapporter/MIT%20nitramine_report_final_July2011.pdf (accessed 1

850 August 2017)

851 Shah, A.D., Liu, Z.-Q., Salhi, E., Hofer, T., Werschkun, B. and von Gunten, U. (2015) Formation of

852 disinfection by-products during ballast water treatment with ozone, chlorine, and peracetic acid: influence of

PT
853 water quality parameters. Environmental Science: Water Research & Technology 1(4), 465-480.

RI
854 Soltermann, F., Abegglen, C., Götz, C. and von Gunten, U. (2016) Bromide sources and loads in Swiss

855 surface waters and their relevance for bromate formation during wastewater ozonation. Environ. Sci.

SC
856 Technol. 50(18), 9825-9834.

U
857 Soltermann, F., Abegglen, C., Tschui, M., Stahel, S. and von Gunten, U. (2017) Options and limitations for
AN
858 bromate control during ozonation of wastewater. Water Res. 116, 76-85.

859 Staehelin, J., Buhler, R.E. and Hoigné, J. (1984) Ozone decomposition in water studied by pulse-radiolysis
M

860 .2. OH and HO4 as chain intermediates. J. Phys. Chem. 88(24), 5999-6004.
D

861 Staehelin, J. and Hoigné, J. (1982) Decomposition of ozone in water: rate of initiation by hydroxide ions and
TE

862 hydrogen peroxide. Environ. Sci. Technol. 16(10), 676-681.

EP

863 Stalter, D., Magdeburg, A. and Oehlmann, J. (2010a) Comparative toxicity assessment of ozone and

864 activated carbon treated sewage effluents using an in vivo test battery. Water Res. 44(8), 2610-2620.
C

865 Stalter, D., Magdeburg, A., Quednow, K., Botzat, A. and Oehlmann, J. (2013) Do Contaminants Originating
AC

866 from State-of-the-Art Treated Wastewater Impact the Ecological Quality of Surface Waters? Plos One 8(4).

867 Stalter, D., Magdeburg, A., Weil, M., Knacker, T. and Oehlmann, J. (2010b) Toxication or detoxication? In

868 vivo toxicity assessment of ozonation as advanced wastewater treatment with the rainbow trout. Water Res.

869 44(2), 439-448.

870 Ternes, T.A., Prasse, C., Eversloh, C.L., Knopp, G., Cornel, P., Schulte-Oehlmann, U., Schwartz, T.,

871 Alexander, J., Seitz, W., Coors, A. and Oehlmann, J. (2017) Integrated Evaluation Concept to Assess the
 ACCEPTED MANUSCRIPT
872 Efficacy of Advanced Wastewater Treatment Processes for the Elimination of Micropollutants and

873 Pathogens. Environ. Sci. Technol. 51(1), 308-319.

874 Tricker, A.R. and Preussmann, R. (1991) Carcinogenic N-nitrosamines in the diet: occurrence, formation,

875 mechanisms and carcinogenic potential. Mutation Research/Genetic Toxicology 259(3), 277-289.

PT
876 U.S. EPA (2006) National primary drinking water regulations: Stage 2 disinfectants and disinfection

877 byproducts rule; Final rule. U.S. Environmental Protection Agency. 40 CFR Parts 9, 141, and 142. Fed.

RI
878 Regist., 71(2): 388–493.

SC
879 von Gunten, U. (2003) Ozonation of drinking water: Part II. Disinfection and by-product formation in

880 presence of bromide, iodide or chlorine. Water Res. 37(7), 1469-1487.

881
U
von Gunten, U. and Hoigné, J. (1994) Bromate formation during ozonization of bromide-containing waters:
AN
882 Interaction of ozone and hydroxyl radical reactions. Environ. Sci. Technol. 28(7), 1234-1242.
M

883 von Gunten, U., Salhi, E., Schmidt, C.K. and Arnold, W.A. (2010) Kinetics and Mechanisms of N-

884 Nitrosodimethylamine Formation upon Ozonation of N,N-Dimethylsulfamide-Containing Waters: Bromide

D

885 Catalysis. Environ. Sci. Technol. 44(15), 5762-5768.

TE

886 von Sonntag, C. and von Gunten, U. (2012) Chemistry of ozone in water and wastewater treatment, IWA
EP

887 Publishing, London.

888 WHO (2008) N-Nitrosodimethylamine in Drinking water.

C

889 http://www.who.int/water_sanitation_health/dwq/chemicals/ndma_2add_feb2008.pdf (accessed 1 August

AC

890 2017)

891 WHO (2011) Guidelines for drinking-water quality, 4th edition.

892 http://apps.who.int/iris/bitstream/10665/44584/1/9789241548151_eng.pdf (accessed 1 August 2017)

893 Wittmer, A., Heisele, A., McArdell, C.S., Böhler, M., Longree, P. and Siegrist, H. (2015) Decreased UV

894 absorbance as an indicator of micropollutant removal efficiency in wastewater treated with ozone. Water Sci.

895 Technol. 71(7), 980-985.

 ACCEPTED MANUSCRIPT
896 Wunderlin, P., Joss, A. and Fleiner, J. (2017) Elimination von Spurenstoffen durch granulierte Aktivkohle

897 (GAK) Filtration: Grosstechnische Untersuchungen auf der ARA Bülach-Furt, Zwischenbericht (accessed

898 1.8.2017).

899 https://www.micropoll.ch/fileadmin/user_upload/Redaktion/Dokumente/01_Berichte/02_Technische_Verfah

900 ren/03_Aktivkohle/GAK_Bu%CC%88lach_Zwischenbericht_27032017_Final.pdf

PT
901 Zietzschmann, F., Mitchell, R.L. and Jekel, M. (2015) Impacts of ozonation on the competition between

organic micro-pollutants and effluent organic matter in powdered activated carbon adsorption. Water Res.

RI
902

903 84, 153-160.

SC
904 Zietzschmann, F., Stutzer, C. and Jekel, M. (2016) Granular activated carbon adsorption of organic micro-

905 pollutants in drinking water and treated wastewater - Aligning breakthrough curves and capacities. Water

906 Res. 92, 180-187.

U
AN
907 Zimmermann, S.G., Wittenwiler, M., Hollender, J., Krauss, M., Ort, C., Siegrist, H. and von Gunten, U.
M

908 (2011) Kinetic assessment and modeling of an ozonation step for full-scale municipal wastewater treatment:

909 Micropollutant oxidation, by-product formation and disinfection. Water Res. 45(2), 605-617.
D
TE

910 Zonja, B., Pérez, S. and Barceló, D. (2016) Human Metabolite Lamotrigine-N2-glucuronide Is the Principal

911 Source of Lamotrigine-Derived Compounds in Wastewater Treatment Plants and Surface Water. Environ.
EP

912 Sci. Technol. 50(1), 154-164.

913
C
AC
 ACCEPTED MANUSCRIPT
The abatement of around 200 micropollutants during biological treatment and ozonation

Micropollutants removal during GAC filtration before and after ozonation was compared

The fate of 19 ozonation products was investigated in various post-treatments

Bromate and NDMA levels in the receiving surface water were below critical values

PT
RI
U SC
AN
M
D
TE
C EP
AC